Abstract: Rail transit is a transportation system that provides urban public passenger services. Its risk management is an arduous and complicated system engineering. In order to ensure the safety of urban rail transit system operation and improve the ability of management departments to respond to emergencies, it is necessary to evaluate the risk status of rail transit operations and formulate corresponding control measures from the aspects of comprehensive application of technology and scientific decision-making. Based on the TEI@I methodology, this paper proposes a delay duration prediction model. The delay time is predicted by establishing a Bayesian network model. The statistical distribution model such as lognormal, Weibull and Gamma distribution is used to verify the prediction result of delay time. Based on the results, a prediction model of subway delay time is constructed. This statistical analysis method is used to calculate the probability of occurrence of the risk event, so as to analyze the risk status of the rail transit system. An empirical study on the Beijing rail transit system shows that the urban rail transit system in Beijing has a good operational status and the possibility of risk events is small. Among all types of risk events, driving accidents are most likely to occur, usually caused by a variety of factors, and the relevant departments should pay enough attention.

Key words: rail transit, risk management, system engineering, Bayesian network, TEI@I